One additional comment, if I may. Anyone who engages in science/religion debates has to have some notion of how strong one’s evidence must be before it becomes irrational to resist it. But judgments of rationality are notorious for being fraught with vagueness and imprecision. And yet we have no choice but to make them, and we do have some clear examples to work with. It would be wonderful if we had a reliable algorithmic procedure that could take all our experience and indubitable propositions as inputs, and produce as an output a conclusion about the rationality of some proposition given the input evidence. But that is a pipe dream. We have to learn to proceed without it, and to just put up with all the vagueness and imprecision.